Shortage of Powerful AI Chips Exposes Bottleneck in Industry Supply Chain
The exponential growth of the AI industry and the rapidly rising demand has exposed a significant bottleneck in the supply chain of powerful chips required for developing and deploying AI models. This shortage is affecting both small and large businesses, including leading platforms in the AI industry.
Industry analysts predict that the situation won’t see any improvement for at least a year or more. The latest sign comes from Microsoft’s annual report, which highlights the bottleneck in the supply chain as a potential risk factor for investors. Microsoft emphasizes the importance of expanding server capacity and data center locations to meet the increasing demand for AI services.
A key component in AI infrastructure, graphics processing units (GPUs), are in short supply. GPUs are vital for running the complex calculations involved in AI training and deployment. Microsoft also identifies other factors influencing GPU production, such as networking supplies, predictable energy, and the availability of land for building.
Microsoft is not the only tech company suffering from this bottleneck. OpenAI CEO Sam Altman testified before the Senate in May, mentioning that ChatGPT was struggling to keep up with the high number of requests due to the shortage of GPUs.
The shortage of chips may remind some of the scarcity of popular consumer electronics during the pandemic, where gaming enthusiasts paid inflated prices for PC graphics cards and game consoles. However, the current shortage is different. It is not driven by disrupted supplies or consumer-focused GPUs but by the sudden demand for ultra-high-end GPUs capable of handling advanced functions like training AI models.
Nvidia, a trillion-dollar company dominating approximately 84% of the discrete GPU market, is expected to benefit greatly from this surge in demand. An investment research note by Moody’s Investors Service predicts that Nvidia will experience unparalleled revenue growth in the near future, with its data center business outperforming rivals AMD and Intel combined.
In response to the growing demand, AMD plans to introduce its AI chip later this year. During their earnings call, AMD CEO Lisa Su expressed strong consumer interest in their AI solutions across the board and highlighted significant progress made by the company.
The shortage of powerful AI chips has highlighted a significant bottleneck in the industry’s supply chain. While companies like Nvidia are expected to thrive, the scarcity of key components poses a challenge for businesses relying on high-performance AI models. The industry will need to address this issue to meet the surging demand and ensure the continued advancement of AI technologies.